Response surface optimization for recovery of polyphenols and carotenoids from leaves of Centella asiatica using an ethanol‐based solvent system

Abstract Response surface methodology has been used to optimize the extraction conditions for total phenolics and carotenoids from leaves of Centella asiatica. Solvent concentration (30%–100%), extraction temperature (30–60°C), and extraction time (30–90 min) were used as the independent variables. A second‐order polynomial model produced a satisfactory fitting of the experimental data with regard to total phenolics (R 2 = 84.75%, p < 0.004) and carotenoid (R 2 = 78.74, p < 0.019) contents. The optimum extraction conditions of ethanol concentration, extraction temperature, and extraction time for phenolics were 6.1%, 70.2°C, and 110.5 min and for carotenoids, the optimum parameters were 100%, 70.2°C, and 110.5 min, respectively. The optimal predicted contents for total phenolics (9.03 mg Gallic Acid Equivalent (GAE)/g DW) and carotenoid (8.74 mg/g DW) values in the extracts were agreed with the experimental values obtained with optimum extraction conditions for each response, and also they possess significantly higher total antioxidant capacity.

Previous studies have also shown that antioxidative activities of different parts of C. asiatica and the phenolic compounds have been suggested as the major contributors to the antioxidative and therapeutic activities (Zainol, Abd-Hamid, Yusof, Muse, 2003). Further, it has been found that bioactive molecules present in C. asiatica can be used as active ingredients for the development of new drugs and natural health products (Pittella, Dutra, Junior, Lopes, Barbosa, 2009).
There is a current trend in investigating natural dietary sources of antioxidants such as green leafy vegetables for the formulation of value-added functional food and nutraceutical ingredients.
Extraction is the initial and most vital step in the recovery and purification of bioactive compounds from plant sources (Prasad et al., 2011). According to Gan and Latiff (2011), many factors such as solvent concentration, extraction temperature, solvent-to-solid ratio, and extraction duration may significantly influence the extraction efficiency and bioactive concentration (Gan & Latiff, 2011). Therefore, it is necessary to optimize the extraction conditions to obtain the highest bioactive recovery. There are investigations showing the antioxidant potential of C. asiatica but none of these explained the optimum extraction conditions for the extraction of bioactives for the use in nutraceutical or pharmaceutical applications. Response surface methodology (RSM) is a widely used tool to evaluate the effects of multiple factors and their interactions in one or more response variables. RSM, nowadays, is one of the most popular optimization techniques in the area of food science and technology and has been applied for the extraction of antioxidant bioactives from a number of dietary sources including Zingiber officinale , olive leaves (Sahin & Samli, 2013), Ipomoea batatas leaves (Song, Li, Liu, & Zhang, 2011), Brassica napus (Wang & Liu, 2009), and Inga edulis leaves (Silva, Pompeu, Larondelle, & Rogez, 2007). There are no studies reported for the optimization of the extraction conditions for polyphenols and carotenoids from leaves of C. asiatica. Therefore, the objective of the present study was to investigate the optimum extraction conditions for C. asiatica leaves to obtain the highest polyphenols and carotenoid content. The findings would be much helpful for the functional foods and nutraceutical industries for the recovery of bioactives from this valuable herb.

| Plant materials
Centella asiatica leaves (type G 1 -"heen gotukola") were collected from home gardens in Makandura area of Sri Lanka, and cleaned edible portions of this leaves were oven-dried at 48°C for 48 hr, ground into powder using a blender, and were stored at −18°C until use. Voucher specimens of the samples have been deposited in the herbarium of the Department of Food Science and Technology of the Wayamba University of Sri Lanka.

| Chemicals
Gallic acid and ethanol were purchased from Sigma-Aldrich, St. Louis, MO, the USA, through Analytical Instrument Pvt Ltd, Colombo, Sri Lanka. All other chemicals used were of analytical grade.

| Preparation of extracts
One gram of air-dried and ground leaf sample was placed in a conical flask with 20 ml aqueous ethanol (1:20 solid/liquid ratio) at desired concentrations, and extraction was carried out for using a rotary shaker (Unimax 1010; Heidolph, Kelheim, Germany) at 400 rpm, at specified temperature as dictated by the experimental design. The response surface optimization procedure was designed based on a three-factor inscribed central composite design (CCD) consisting of aqueous ethanol (30%-100%), extraction temperature (30-60°C), and extraction time (30-90 min) as shown in Table 1. The extracts were then filtered through a filter paper (Whatman No. 42;Whatman Paper Ltd,Maidstone,UK), and the filtrates were stored at −18°C until used for the determination of total polyphenols and carotenoid contents.

| Determination of total polyphenol content
The total polyphenol content was determined using Folin-Ciocalteu assay (Singleton, Orthofer, & Lamuela-Raventos, 1999) with some modification, as described by Gunathilake, Yu, and Rupasinghe (2014) and Gunathilake (2012). About 0.5 ml of leaf extract and 0.1 ml of Folin-Ciocalteu reagent (0.5N) were mixed and incubated at room temperature (30°C) for 15 min at dark. Sodium carbonate (7.5%, 25 ml) was added and incubated for further 2 hr at dark. Absorbance was measured at 760 nm using UV/ VIS spectrometer (Optima, SP-3000, and Tokyo, Japan). Gallic acid was used to prepare standard curve, and the concentration of total polyphenols was expressed as mg of gallic acid equivalents (GAE) per gram dry weight.

| Total carotenoid content
The carotenoid content was analyzed according to the method described by Şükran, Gunes, and Sivaci (1998)

| Determination of total antioxidant capacity
The total antioxidant capacity of leaf extracts was analyzed according to the method described by Prieto, Pineda, and Aguilar (1999) with some modifications of Gunathilake and Ranaweera (2016).
Briefly, 0.3 ml leaf extract and 3 ml reagent solution (0.6 M sulphuric acid, 28 mM sodium phosphate, and 4 mM ammonium molybdate) were incubated at 95°C for 90 min, and then, the solution was cooled to room temperature (30 ± 2°C), and the absorbance of each solution was measured at 695 nm spectrophotometrically against a blank. The antioxidant capacity was expressed as ascorbic acid equivalents (AAE).

| Experimental design
Optimization of extraction parameters of phenolics from C. asiatica leaves was done using RSM.
Influence of three independent variables, ethanol concentration, extraction temperature, and extraction time and the response variables were total phenolic, and total carotenoid contents were studied. A three-factor inscribed CCD was used to identify the relationship existing between the response functions and the process variables, as well as to determine those conditions that optimized the extraction process of total phenolics and carotenoid contents of the extracts. The independent variables and the range studied were ethanol concentration (30-100%), temperature (30-60°C), and extraction time (30-90 min), and solid-to-liquid ratio was maintained at 1:20. The selection and range of these three factors were based on previous studies. Each variable to be optimized was coded at three levels 1, 0, +1 (Table 1). According to the design used, twenty randomized experiments including six replicates as the center points were assigned based on CCD and the values of independent process variables considered, as well as measured total phenolic content and carotenoid content, are given in Table 2. fit the mathematical models of the experimental data that aimed to optimize the overall region for both response variables. A secondorder polynomial model was used to predict the response variables as appeared below: where Y is the predicted dependent variable; β 0 is a constant that fixes the response at the central point of the experiment; β 1 , β 2 , and β 3 are the regression coefficients for the linear effect terms; 2 1 , 2 2 , and 2 3 are the quadratic effect terms; and β 1 β 2 , β 1 β 3 , and β 2 β 3 are the interaction effect terms, respectively. X1, X2, and X3 are the independent variables (Table 1). The adequacy of the model was predicted through the regression analysis (R 2 ) and the ANOVA analysis. The relationship between the independent variables and the response variables (polyhenols and carotenoids) was demonstrated by the response surface plots. Multiple graphical and numerical optimizations of the experimental data were done to identify the optimum extraction conditions to achieve the maximum recovery of polyphenols and carotenoids.

| Statistical design
Verification of predicted extraction conditions that would give higher levels of polyphenols and carotenoids was determined based on the best extractions conditions obtained with RSM.

| RE SULTS AND D ISCUSS I ON
Twelve morphotypes of C. asiatica have been recorded according to their morphological and morphometric characters in Sri Lanka and among the types available, types G1 and G2 known as "heen gotukola" which is having smaller leaves and are more popular among the local community as highly nutritious types. Types G8 and G12 known as "giant gotukola" have very large leaves, and they contained higher amount of β-carotene and lutein (Chandrika, Salim, Wijepala, Perera, & Goonetilleke, 2011). The presence of various phenolic and carotenoid bioactives such as triterpene saponins, asiaticoside, numerous caffeic acid derivatives, and flavonoids in C. asiatica is believed to be responsible for health benefits associated with this leafy vegetable (Chippada & Vangalapati, 2011).
An optimization of extraction conditions for the recovery of total polyphenols and carotenoids from C. asiatica was conducted using RSM. The extraction efficiency of these bioactive constituents was influenced by extraction solvent properties, extraction time, and extraction temperature (Alothman, Bhat, & Karim, 2009 Table 2. The obtained data were used for the prediction of an optimum set of extraction parameters from the leaf extract with high polyphenols and carotenoids. The concentration of polyphenols and carotenoids in the extracts was employed in a multiple regression analysis, performed using RSM to fit the second-order polynomial equations is given in Table 3 for polyphenols and carotenoids, respectively. The "fitness" of the model was studied through the lack-of-fit test (p ≤ 0.05), which indicated the adequacy of models to accurately predict the variation (Kong, Ismail, Tan, Prasad, & Ismail, 2010). The quality of fit to the second-order polynomial models for leaf extracts of C. asiatica was established based on the coefficients of determination (70%> R 2 ), regression p-value (p ≤ 0.1), and lack of fit (p ≤ 0.05) indicating that the models could be used to predict the responses.
The software generated the estimated regressions coefficients for quadratic equations as appeared in Table 3.

| Model fitting of parameters based on total phenolic and carotenoid content
For RSM, the levels of independent variables for the extraction of total polyphenols and carotenoids were selected based on the literature. The experimental design and corresponding response data are presented in

| Effect of extraction parameters on total phenolic content
Response surfaces were used to illustrate the effects of solvent concentration, extraction time, and the temperature on the responses (Figures 1 and 2). The responses demonstrated that the ethanol concentration, extraction temperature, and the duration of the ex- for C. asiatica ( Figure 1). As the extraction and isolation of polyphenols depend greatly on the polarity of the extraction solvent, the use of a pure solvent may not be effective for the separation of polyphenols from plant materials as described in Prasad et al. (2011). Further, according to Lang and Wai (2006), water is acting as the plant swelling agent, while ethanol may disrupt the bonding between the solutes and plant matrices. This is indicating that the mixture of water and ethanol or "aqueous ethanol" as solvent agent exhibited the best performance to extract polyphenols from plant sources. Therefore, a combination of alcohol with water seems more effective in extracting polyphenols. This is consistent with several earlier findings which convey that polyphenols are more extractable in polar solvents as compared to nonpolar ones (Hayouni et al., 2007;Prasad et al., 2011).
Regarding extraction temperature on total polyphenols, the recovery of phenolics was increased considerably when the extraction temperature was increased to 60°C, while the % ethanol maintained at a low level (Figure 1c and d). Results showed that at lower solvent concentration (30%), the use of higher extraction temperature (60°C) and extraction time (90 min) increased the extractable phenolics from 3.31 to 4.13 mg GAE/g DW, compared with the use of lower extraction temperature (30°C) and extraction time (30 min).
This could be due to the increase in the solubility of polyphenols, diffusion rate, mass transfer rate, extraction rate, and reduced solvent viscosity and surface tension at higher temperatures and solvent polarities which could improve the polyphenol extractability (Richter et al., 1996). The extraction time was another important parameter in the extraction procedure for bioactives in many previous studies Sahin & Samli, 2013). However, the results showed that extraction time did not have a significant effect on the polyphenol extraction from C. asiatica leaves at p ≤ 0.05 level.

| Effect of extraction parameters on carotenoid content
Various solvent systems have been used for carotenoid extraction, and ethanol is also a good solvent that can be used for carotenoid extraction (Ofori-Boateng & Lee, 2013) and the extraction is highly TA B L E 3 Estimated regression coefficients of the second-order polynomial equations for RSM analysis of total polyphenols and carotenoid extraction (uncoded) from Centella asiatica influenced by extractions variables including solvent concentration, extraction temperature, and time (Wang & Liu, 2009). Many researchers have used anon-polar solvent for carotenoid extraction, petroleum ether/acetone (1/1) for rapeseed (Wang & Liu, 2009), and hexane/acetone/alcohol (2/1/1) for lycopene (Kaur, Wani, Oberoi, & Sogi, 2008

| Optimization of polyphenols and carotenoids and verification of the model
Optimum process parameters achieved by maximizing total phenolics and carotenoid contents. During the optimization stage, the desirability function of the MINITAB statistical software is used to obtain the best compromise of the two responses with the weights of all 1.0. As shown in Table 4, the predicted optimal ethanol concentration, extraction temperature and extraction time were developed for maximizing the both responses, and they were 37%, 70.20°C, and 110.5 min for phenolics and 100%, 70.20°C, and 110.5 min for carotenoids, respectively. For these optimum extraction conditions, the corresponding predicted response values for phenolics and carotenoids were 4.71 mg GAE/g DW and 3.55 mg/g DW, respectively. An experiment was run by the recommended optimum conditions for two responses, phenolics, and carotenoids. More interestingly, in this study, the values obtained experimentally for both response variables are near to the predicted values, indicating a satisfactory model. The experimental values for total phenolics were 4.58 ± 0.44 mg GAE g extract and 3.28 ± 0.39 mg/g DW carotenoids, and no significant difference (p < 0.05) was found between the experimental and predicted values of the extractable phenolics and carotenoids from leaves of C. asiatica extract. Further, the extracts prepared with the optimum extraction conditions showed significantly higher (p < 0.05) total antioxidant capacity compared with the extract prepared with 100% ethanol, 30°C, and 30 min extraction conditions. Therefore, the data confirm the validity of the optimized model.

| CON CLUS IONS
An ethanol-based extraction technique was applied for the extraction of polyphenols and carotenoid compounds from C. asiatica leaves and optimized by response surface methodology. The results showed that the extraction conditions including solvent concentration, extraction temperature, and extraction time markedly influenced the yields of total phenolics and total carotenoids of the C. asiatica extracts. Overall, extraction of polyphenols prefers low ethanol concentration, higher temperature, and longer ex-